Learning Lightweight Dynamic Kernels With Attention Inside via Local-Global Context Fusion

doi:10.1109/TNNLS.2022.3217301

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 平行智能技术与系统团队

	Learning Lightweight Dynamic Kernels With Attention Inside via Local-Global Context Fusion
	Tian, Yonglin1 ; Shen, Yu1 ; Wang, Xiao1 ; Wang, Jiangong1 ; Wang, Kunfeng2 ; Ding, Weiping 3; Wang, Zilei 4; Wang, Fei-Yue1
发表期刊	IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
ISSN	2162-237X
	2022-11-14
页码	15
通讯作者	Wang, Fei-Yue(feiyue.wang@ia.ac.cn)
摘要	Traditional convolutional neural networks (CNNs) share their kernels among all positions of the input, which may constrain the representation ability in feature extraction. Dynamic convolution proposes to generate different kernels for different inputs to improve the model capacity. However, the total parameters of the dynamic network can be significantly huge. In this article, we propose a lightweight dynamic convolution method to strengthen traditional CNNs with an affordable increase of total parameters and multiply-adds. Instead of generating the whole kernels directly or combining several static kernels, we choose to "look inside ", learning the attention within convolutional kernels. An extra network is used to adjust the weights of kernels for every feature aggregation operation. By combining local and global contexts, the proposed approach can capture the variance among different samples, the variance in different positions of the feature maps, and the variance in different positions inside sliding windows. With a minor increase in the number of model parameters, remarkable improvements in image classification on CIFAR and ImageNet with multiple backbones have been obtained. Experiments on object detection also verify the effectiveness of the proposed method.
关键词	Attention inside kernels dynamic convolution global context local context
DOI	10.1109/TNNLS.2022.3217301
收录类别	SCI
语种	英语
资助项目	Key-Area Research and Development Program of GuangdongProvince[2020B090921003] ; Key Research andDevelopment Program of Guangzhou[202007050002] ; Natural Science Key Foundation of Jiangsu Education Department[21KJA510004] ; Intel Collaborative Research Institute forIntelligent and Automated Connected Vehicles (ICRI-IACV) ; National Natural Science Foundation of China[62076020] ; National Natural Science Foundation of China[U1811463] ; National Natural Science Foundation of China[61976120] ; National Natural Science Foundation of China[62173329]
项目资助者	Key-Area Research and Development Program of GuangdongProvince ; Key Research andDevelopment Program of Guangzhou ; Natural Science Key Foundation of Jiangsu Education Department ; Intel Collaborative Research Institute forIntelligent and Automated Connected Vehicles (ICRI-IACV) ; National Natural Science Foundation of China
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Computer Science, Hardware & Architecture ; Computer Science, Theory & Methods ; Engineering, Electrical & Electronic
WOS记录号	WOS:000886698500001
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：2[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/51279
专题	多模态人工智能系统全国重点实验室_平行智能技术与系统团队
通讯作者	Wang, Fei-Yue
作者单位	1.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China 2.Beijing Univ Chem Technol, Coll Informat Sci & Technol, Beijing 100029, Peoples R China 3.Nantong Univ, Sch Informat Sci & Technol, Nantong 226019, Peoples R China 4.Univ Science & Technol China, Natl Engn Lab Brain inspired Intelligence Technol, Hefei 230027, Peoples R China
第一作者单位	中国科学院自动化研究所
通讯作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Tian, Yonglin,Shen, Yu,Wang, Xiao,et al. Learning Lightweight Dynamic Kernels With Attention Inside via Local-Global Context Fusion[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2022:15.
APA	Tian, Yonglin.,Shen, Yu.,Wang, Xiao.,Wang, Jiangong.,Wang, Kunfeng.,...&Wang, Fei-Yue.(2022).Learning Lightweight Dynamic Kernels With Attention Inside via Local-Global Context Fusion.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,15.
MLA	Tian, Yonglin,et al."Learning Lightweight Dynamic Kernels With Attention Inside via Local-Global Context Fusion".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2022):15.